pandas-eda

pandas EDA

Big Data? Machine Learning?

You work with your data -
Manipulating it, merging, pivot and more.

But what happened in between? Did you get a lot of nans? Maybe duplicated values?
You constantly need to check your data status,
But you cannot do values_counts, isna and all those stuffs every second...

pandas_eda is an Exploratory Data Analysis tool that will show you status and frequent values for each column!
You will be focused on what you have on the spot.

Demo is the best way to understand:

install:

pip install pandas_eda

usage:

import pandas as pd
import pandas_eda
from time import sleep


# dummy data
df = pd._testing.makeMixedDataFrame()
# or load your data with pd.read_excel('whatever.xlsx')


# show original data. will pop up a web application
df.eda()  # can use this at debug mode too!

# that's it. but you can do more...


# manipulation #1
df.A += 10
process = df.eda(title='take 2')  # yes, you can open multiple EDA windows!
# process.kill()


# manipulation #2
df.loc[df.B==1, 'A'] -= 30
df.eda()


# no need at jupyter or debug mode...
# ending the script will end the eda too, so delaying the exit. 
sleep(600)  

headless mode

import pandas as pd
import pandas_eda

# disabling the wrapping of a long table at the print...
pd.options.display.expand_frame_repr = False
pd.options.display.max_columns = 0
pd.options.display.min_rows = 20

df = pandas_eda._testing.generate_fake_table(samples=1_000)
eda = df.eda(cli_mode=True)

print('\n\n *** column statistics *** ')
print(eda.get_columns_statistics())
print('\n\n *** frequent values *** ')
print(eda.get_frequent_values())

note:
If you're running on remote machine, the eda will be opened on the remote...

alternatives:
After starting this tool I've found 2 cool alternatives:

• sweetviz
  Has a nice interactive report.
• mito
  Great for new table that needs also cleaning.
  Works only at jupyter.